Objective  To develop a deep learning method for small sample multi-omics data using attention mechanism and Meta-learning for the establishment of autism diagnosis model.

Methods  MLAN (Meta-learning based attentive network) consisting of the omics feature pre-reduction module, the multi-omics data fusion and feature learning module, and the parameter optimization module was designed. Firstly, differential expression analysis was performed on high-dimensional multi-omics data to preliminarily screen out unimportant features. Secondly, a multi-channel attention mechanism was used to learn the importances of every set of omics data and to realize data fusion, and a two-layer fully connected network was constructed to further extract latent features and realize the diagnosis task. Finally, the Meta-learning algorithm Reptile was used to optimize the initial parameters of the above model to obtain the optimal parameters. A total of 58 children’s saliva samples were collected, including 21 children diagnosed with autism, 12 children with social disorders, and 25 healthy controls, and the protein and metabolomics data were detected by mass spectrometry. All data were randomly divided into training set and test set by 4 ∶ 1, and the training set was divided into training data and validation data in the same way for model training and validation. The test set was used for the final evaluation of the model effect. Five baseline models and three ablated models were constructed and evaluated along with MLAN based on metrics including multi-classification accuracy, F1-macro and F1-weighted scores.

Results  The constructed multi-classification autism diagnosis model MLAN achieved multi-classification accuracy, F1-macro and F1-weighted scores of 0.850±0.066, 0.817±0.103 and 0.834±0.087. The values of all three indicators were better than those of baseline models and the ablated models.

Conclusion  The proposed MLAN can effectively deal with heterogeneous multi-omics data with small samples and achieve good results, which is expected to provide assistance for the clinical diagnosis of autism.

1.Volkmar FR. Encyclopedia of autism spectrum disorders[M]. Cham: Springer International Publishing, 2021: 1-691.

2.Zhou H, Xu X, Yan W, et al. Prevalence of autism spectrum disorder in China: a nationwide multi-center population-based study among children aged 6 to 12 years[J]. Neurosci Bull, 2020, 36(9): 961-971. DOI: 10.1007/s12264-020-00530-6.

3.Mottron L, Bzdok D. Autism spectrum heterogeneity: fact or artifact?[J]. Mol Psychiatry, 2020, 25(12): 3178-3185. DOI: 10.1038/s41380-020-0748-y.

4.李雨田, 郭小芳, 梁连娇. 孤独症谱系障碍早期诊断与早期干预模式探讨[J]. 中国现代药物应用, 2015, 9(14): 260-262. DOI: 10.14164/j.cnki.cn11-5581/r.2015.14.180.

5.李丹丹, 王鸿武, 吴毅, 等. 孤独症谱系障碍筛查工具的演变和应用研究进展[J]. 环境与健康杂志, 2025, 42(2): 182-187. [Li DD, Wang HW, Wu Y, et al. Evolution and application of screening tools for autism spectrum disorder: a review of recent studies[J]. Environment & Health, 2025, 42(2): 182-187.] DOI: 10.16241/j.cnki.1001-5914.2025.02.017.

6.Tomiyama S, Yoshida K, Tani H, et al. Pharmacological treatment of autism spectrum disorder: a systematic review of treatment guidelines[J]. Pharmacopsychiatry, 2025, 58(3): 100-116. DOI: 10.1055/a-2514-4452.

7.Dawson G, Rieder AD, Johnson MH. Prediction of autism in infants: progress and challenges[J]. Lancet Neurol, 2023, 22(3): 244-254. DOI: 10.1016/S1474-4422(22)00407-0.

8.赵宇曦. 基于蛋白质组学和代谢组学的自闭症血浆生物标志物研究[D]. 广东深圳: 深圳大学, 2020.

9.徐宇航. 基于图神经网络和MRI影像的自闭症辅助诊断方法研究[D]. 成都: 电子科技大学, 2024. DOI: 10.27005/d.cnki.gdzku.2024.000861.

10.韩俊林, 林玉梅, 林堃, 等. 孤独症谱系障碍儿童的肠道菌群特点及其对患儿行为的影响[J]. 中外医学研究, 2023, 21(34): 63-67. [Han JL, Lin YM, Lin K, et al. Characteristics of intestinal flora in children with autism spectrum disorder and its influence on children's behavior[J]. Chinese and Foreign Medical Research, 2023, 21(34): 63-67.] DOI: 10.14033/j.cnki.cfmr.2023.34.016.

11.张彩云, 候芳, 陈艳琳, 等. 家系全外显子测序揭示自闭症谱系障碍中的关键遗传变异和生物通路[C]//深圳市康复医学会, 香港职业治疗学院, 大湾区康复医学会, 等. 2024年深圳国际康复论坛（第二十一届）优秀摘要集, 2024: 7-8. DOI: 10.26914/c.cnkihy.2024.026434.

12.龙智平, 王帆. 多组学整合分析的设计及统计方法在肿瘤流行病学研究中的应用[J]. 中华流行病学杂志, 2020, 41(5): 788-793. [Long ZP, Wang F. Study design and statistical methods used for integrative analysis on multi-omics in cancer epidemiology[J]. Chinese Journal of Epidemiology, 2020, 41(5): 788-793.] DOI: 10.3760/cma.j.cn112338-20190624-00461.

13.Choi JM, Chae H. moBRCA-net: a breast cancer subtype classification framework based on multi-omics attention neural networks[J]. BMC Bioinformatics, 2023, 24(1): 169. DOI: 10.1186/s12859-023-05273-5.

14.Hu J, Shen L, Sun G. Squeeze-and-excitation networks[C]. Proceedings of the IEEE conference on computer vision and pattern recognition, 2018: 7132-7141.

15.Nichol A, Achiam J, Schulman J. On first-order Meta-learning algorithms[EB/OL]. (2018-04-04) [2025-04-25]. https://doi.org/10.48550/arXiv.1803.02999.

16.American Psychiatric Association. Diagnostic and statistical manual of mental disorders, DSM-5-TR[M]. Arlington, VA: American Psychiatric Publishing, 2022: 56-68.

17.Krug DA, Arick J, Almond P. Behavior checklist for identifying severely handicapped individuals with high levels of autistic behavior[J]. J Child Psychol Psychiatry, 1980, 21(3): 221-229. DOI: 10.1111/j.1469-7610.1980.tb01797.x.

18.Schopler E, Reichler RJ, DeVellis RF, et al. Toward objective classification of childhood autism: Childhood Autism Rating Scale (CARS)[J]. J Autism Dev Disord, 1980, 10(1): 91-103. DOI: 10.1007/BF02408436.

19.谢堃, 陈雯, 陈锋, 等. 一种唾液采集装置: 中国专利, 221770063U[P]. 2024-09-27.

20.Hu S, Loo JA, Wong DT. Human saliva proteome analysis[J]. Ann Ny Acad Sci, 2007, 1098(1): 323-329. DOI: 10.1196/annals.1384.015.

21.Mota FSB, Nascimento KS, Oliveira MV, et al. Potential protein markers in children with Autistic Spectrum Disorder (ASD) revealed by salivary proteomics[J]. Int J Biol Macromol, 2022, 199: 243-251. DOI: 10.1016/j.ijbiomac.2022.01.011.

22.Dame ZT, Aziat F, Mandal R, et al. The human saliva metabolome[J]. Metabolomics, 2015, 11: 1864-1883. DOI: 10.1007/s11306-015-0840-5.

23.李圆圆. 基于质谱的微量样品代谢组与蛋白质组新方法研究[D]. 沈阳: 东北大学, 2022. DOI: 10.27007/d.cnki.gdbeu.2022.003435.

24.Wei R, Wang J, Su M, et al. Missing value imputation approach for mass spectrometry-based metabolomics data[J]. Sci Rep, 2018, 8(1): 663. DOI: 10.1038/s41598-017-19120-0.

25.Yang H, Chen R, Li D, et al. Subtype-GAN: a deep learning approach for integrative cancer subtyping of multi-omics data[J]. Bioinformatics, 2021, 37(16): 2231-2237. DOI: 10.1093/bioinformatics/btab109.

26.Shehab M, Abualigah L, Shambour Q, et al. Machine learning in medical applications: A review of state-of-the-art methods[J]. Comput Biol Med, 2022, 145: 105458. DOI: 10.1016/j.compbiomed.2022.105458.

27.Levy S, Duda M, Haber N, et al. Sparsifying machine learning models identify stable subsets of predictive features for behavioral detection of autism[J]. Mol Autism, 2017, 8: 65. DOI: 10.1186/s13229-017-0180-6.

28.Rappoport N, Shamir R. Multi-omic and multi-view clustering algorithms: review and cancer benchmark[J]. Nucleic Acids Res, 2018, 46(20): 10546-10562. DOI: 10.1093/nar/gky889.

29.Wang Y, Wang Z, Yu X, et al. MORE: a multi-omics data-driven hypergraph integration network for biomedical data classification and biomarker identification[J]. Brief Bioinform, 2025, 26(1): bbae658. DOI: 10.1093/bib/bbae658.

30.Ferri C, Hernández-Orallo J, Modroiu R. An experimental comparison of performance measures for classification[J]. Pattern Recognit. Lett, 2009, 30(1): 27-38. DOI: 10.1016/j.patrec.2008.08.010.

31.Ben-Sasson A, Guedalia J, Ilan K, et al. Predicting autism traits from baby wellness records: a machine learning approach[J]. Autism, 2024, 28(12): 3063-3077. DOI: 10.1177/13623613241253311.

32.Jones W, Klaiman C, Richardson S, et al. Eye-tracking-based measurement of social visual engagement compared with expert clinical diagnosis of autism[J]. JAMA, 2023, 330(9): 854-865. DOI: 10.1001/jama.2023.13295.

33.Jiao Y, Chen R, Ke X, et al. Single nucleotide polymorphisms predict symptom severity of autism spectrum disorder[J]. J Autism Dev Disord, 2012, 42(6): 971-983. DOI: 10.1007/s10803-011-1327-5.

34.Ottosson F, Russo F, Abrahamsson A, et al. Unraveling the metabolomic architecture of autism in a large Danish population-based cohort[J]. BMC Med, 2024, 22(1): 302. DOI: 10.1186/s12916-024-03516-7.

35.Szoko N, McShane AJ, Natowicz MR. Proteomic explorations of autism spectrum disorder[J]. Autism Res, 2017, 10(9): 1460-1469. DOI: 10.1002/aur.1803.

36.Guo LY, Wu AH, Wang YX, et al. Deep learning-based ovarian cancer subtypes identification using multi-omics data[J]. BioData Min, 2020, 13(1): 10. DOI: 10.1186/s13040-020-00222-x.

37.Hira MT, Razzaque MA, Angione C, et al. Integrated multi-omics analysis of ovarian cancer using variational autoencoders[J]. Sci Rep, 2021, 11(1): 6265. DOI: 10.1038/s41598-021-85285-4.

38.Shao W, Zuo Y, Shi YY, et al. Characterizing the survival-associated interactions between tumor-infiltrating lymphocytes and tumors from pathological images and multi-omics data[J]. IEEE Trans Med Imaging, 2023, 42(10): 3025-3035. DOI: 10.1109/TMI.2023.3274652.

39.Pan L, Wang X, Liang Q, et al. DEDUCE: Multi-head attention decoupled contrastive learning to discover cancer subtypes based on multi-omics data[J]. Comput Methods Programs Biomed, 2024, 257: 108478. DOI: 10.1016/j.cmpb.2024.108478.

40.Sun Q, Cheng L, Meng A, et al. SADLN: self-attention based deep learning network of integrating multi-omics data for cancer subtype recognition[J]. Front Genet, 2023, 13: 1032768. DOI: 10.3389/fgene.2022.1032768.

41.Wen G, Li L. MMOSurv: Meta-learning for few-shot survival analysis with multi-omics data[J]. Bioinformatics, 2024, 41(1): btae684. DOI: 10.1093/bioinformatics/btae684.

42.Huang HH, Shu J, Liang Y. MUMA: a multi-omics Meta-learning algorithm for data interpretation and classification[J]. IEEE J Biomed Health Inform, 2024, 28(4): 2428-2436. DOI: 10.1109/JBHI.2024.3363081.

43.Finn C, Abbeel P, Levine S. Model-agnostic Meta-learning for fast adaptation of deep networks[C]. International conference on machine learning. PMLR, 2017: 1126-1135.

44.Geschwind DH, Levitt P. Autism spectrum disorders: developmental disconnection syndromes[J]. Curr Opin Neurobiol, 2007, 17(1): 103-111. DOI: 10.1016/j.conb.2007.01.009.

45.Lord C, Elsabbagh M, Baird G, et al. Autism spectrum disorder[J]. Lancet, 2018, 392(10146): 508-520. DOI: 10.1016/S0140-6736(18)31129-2.

46.Heinsfeld AS, Franco AR, Craddock RC, et al. Identification of autism spectrum disorder using deep learning and the ABIDE dataset[J]. Neuroimage Clin, 2018, 17: 16-23. DOI: 10.1016/j.nicl.2017.08.017.

47.He H, Garcia EA. Learning from imbalanced data[J]. IEEE Trans Knowl Data Eng, 2009, 21(9): 1263-1284. DOI: 10.1109/TKDE.2008.239.

48.Wang S, Wang S, Wang Z. A survey on multi-omics-based cancer diagnosis using machine learning with the potential application in gastrointestinal cancer[J]. Front Med (Lausanne), 2023, 9: 1109365. DOI: 10.3389/fmed.2022.1109365.

49.Harrell FE. Regression modeling strategies: with applications to linear models, logistic and ordinal regression, and survival analysis, 2nd edition[M]. Cham: Springer, 2015: 63-102.